Real jobs people get done with AI agents, each tied to its original source, an honest outcome label, a prompt you can edit, and the GitHub repositories involved.
The Manus team describes the context rules it settled on after rebuilding its agent framework four times: keep the prompt prefix cacheable, mask tools instead of removing them, and use files as memory.
OpenHands Agent Canvas runs the OpenHands agent, or Claude Code, Codex, Gemini CLI and other ACP agents, on your machine or in Docker sandboxes, and can trigger them from schedules or webhooks.
godot-mcp lets an MCP client launch the Godot editor, run a project, read its debug output and create scenes, so the agent can check whether its game code works.
Replit Agent builds and browser-tests apps from a description. A widely reported 2025 incident, where it deleted a live database during a code freeze, shows why separate dev and prod data matters.
Gemini CLI is Google’s open-source terminal agent. Simon Willison found it strong on large codebases in 2025; in 2026 Google moved individual users to Antigravity CLI.
Start Cursor agents from the editor, web, Slack or a GitHub comment; each works in its own cloud VM on a branch and returns a pull request with screenshots, videos and logs.
Cognition reports customers giving Devin written migration and testing playbooks, then running many sessions in parallel; its own numbers show where Devin does well and where it needs a human.
B12 CTO Adam Marcus used Claude Code to analyze gigabyte core dumps filling web server disks and traced them to an RSS endpoint reading an endless radio stream.
Matthew Brunelle rebuilt a YouTube Music to OpenSubsonic shim with OpenCode and Kimi K2.6 to compare it with an earlier Claude Code and Opus 4.6 attempt.
Mitchell Hashimoto built Ghostty's unobtrusive macOS update notifications with the Amp agent over 16 sessions, rewriting and reviewing parts by hand before merging.
Peter Eysermans connected Claude Desktop to Playwright and GitHub MCP servers so it could walk through features from tickets, report results and draft reusable tests.
Birgitta Böckeler tried GitHub's spec-kit workflow with Copilot in VS Code on a mid-sized dashboard feature and found the process heavy for the size of the task.
Simon Willison had Codex CLI translate the JustHTML parser into a dependency-free JavaScript library, steering it with about eight prompts over one afternoon.